Dr. Denis

> It does not compile how it is now, see comments in code. > > Could you please add a proper documentation of this controller in docs/userdoc.rst for our sphinx...

> 2 - Most of the controller parameters are defined as read-only. That means that they need to be set when loading the controller manager. Usually, I set up my...

> 3 - The feedforward input could be more generic. Instead of tying it to the reference it could be left as another generic input, which is multiplied by its...

PID controller should not set values on NaN on deactivate. We should change this here. Fixed in ff17e8a – please check.

> @destogl thanks for the fix, but I still wonder what is the correct procedure in on_activate. Should the controller initialize the command to the current state (and ignore any...

Tests are clean on rolling: Hardware Interface: ``` build/hardware_interface/Testing/20240619-1807/Test.xml: 6 tests, 0 errors, 0 failures, 0 skipped build/hardware_interface/test_results/hardware_interface/test_component_interfaces.gtest.xml: 9 tests, 0 errors, 0 failures, 0 skipped build/hardware_interface/test_results/hardware_interface/test_component_parser.gtest.xml: 40 tests, 0...

Merging as agreed in WG Meeting yesterday.

resolved in #1562

@tfoote didn't we resolve this issue in #240

Hi, Dr. Denis here, maintainer of ros2_control. I am wondering, why do you have your implementation of the `Fake Hardware`? Does the generic implementation of (now renamed!) [`Mock Hardware`](https://github.com/ros-controls/ros2_control/blob/master/hardware_interface/src/mock_components/generic_system.cpp) is...